[Proposition] Trusting players

aix
Member
 
Posts: 15
Joined: Fri Jul 22, 2016 18:52
IRC: aix
In-game: aix

[Proposition] Trusting players

by aix » Wed Aug 03, 2016 11:51

It was not so long ago that Minetest server owners had no viable options on the "node protection" front, and servers were badly griefed by pre-pubescent teens trying to gain materials, or, perhaps even worse, intentionally trying to cause damage. This is still the case today, but unlike the 0.3.x times, Minetest is now available on Android, which is exactly the type of device you'd give to your children.
Sure, now we have all sorts of protection mods and tools to prevent or rollback griefing, but it's not as if the problem of griefing lies with Minetest, is it? The problem of griefing lies with the kiddies who see a server in the server list, join it, and cause havoc. We can minimize damage by using, say, areas, but what are we really proving? That we don't trust our players, and that our players do not trust each other. Even using the keyword mod which requires new users to enter a keyword in chat doesn't completely cut down on kiddies. I set it up on my server, and while completely blocking kiddies, it doesn't do anything about the join/leave messages or the terrible spam they cause when they discover they cannot place blocks.

So how do we fight this? Well, I plan on making an "exclusive" server, for players who have proved that they are of sound mind and can grasp the concept of fair play.
Now, I'm stuck with a dilema.
If I leave the server in the global list, kiddies will join.
If I don't, how will people know to join the server?

I guess a solution would be to make a "quiz board", which you can sign up for, get an interview on IRC, then have your name added to a list of "sane" users, but, then we have the obvious problems of staffing this and getting server owners to incorporate this.
MinePass seems like it'd do the job quite well, but I'm not too familiar with how it works yet.

What do you think? What would I need to make a server (or network of servers) that trust their players and that minimize spam by vetting the users prior to them joining?
Lightning Server
Courage is not the absence of fear, but rather the judgment that something else is more important than one's fear.
-Ambrose Redmoon
 

User avatar
maikerumine
Member
 
Posts: 1348
Joined: Mon Aug 04, 2014 14:27
GitHub: maikerumine
In-game: maikerumine
 

CWz
Member
 
Posts: 196
Joined: Tue Dec 24, 2013 17:01
Location: Banana Land
GitHub: chaoswormz
IRC: CWz
In-game: CWz

Re: [Proposition] Trusting players

by CWz » Wed Aug 03, 2016 14:13

I use my blockade mod for my creative server. it blocks new registrations and add a create_user command for adding players.

That is the only way, i can think of that would "raise the player quality".

Unfortunately the rise of the tablet version and the fake tablet versions brought with them hoards of hyperactive scream-o'-trons, and since minetest is drinking up some of that good awful popularity milksludge straight from the liquid sporecloud we will see an increase of stonemen from here on out.
Last edited by CWz on Wed Aug 03, 2016 14:27, edited 1 time in total.
 

aix
Member
 
Posts: 15
Joined: Fri Jul 22, 2016 18:52
IRC: aix
In-game: aix

Re: [Proposition] Trusting players

by aix » Wed Aug 03, 2016 14:19

maikerumine wrote:Remove interact and white list it. ;)

That doesn't stop users joining and messaging everyone.
Also, if I remove shout by default, players can't ask for interact.
Lightning Server
Courage is not the absence of fear, but rather the judgment that something else is more important than one's fear.
-Ambrose Redmoon
 

CWz
Member
 
Posts: 196
Joined: Tue Dec 24, 2013 17:01
Location: Banana Land
GitHub: chaoswormz
IRC: CWz
In-game: CWz

Re: [Proposition] Trusting players

by CWz » Wed Aug 03, 2016 14:29

aix wrote:
maikerumine wrote:Remove interact and white list it. ;)

That doesn't stop users joining and messaging everyone.
Also, if I remove shout by default, players can't ask for interact.


My keyword mod works even without shout privs. also if you want you could also use cheapie's newplayer mod which doesn't use chat but rather a formspec. it's also harder for newbe's to figure out.

Cheapie's new player mod

The whitelist mod only players on the list can join.

My keyword interact mod

a fork of my keyword mod with slightly more advanced feature(not made by me)
 

aix
Member
 
Posts: 15
Joined: Fri Jul 22, 2016 18:52
IRC: aix
In-game: aix

Re: [Proposition] Trusting players

by aix » Wed Aug 03, 2016 14:51

CWz wrote:My keyword mod works even without shout privs. also if you want you could also use cheapie's newplayer mod which doesn't use chat but rather a formspec. it's also harder for newbe's to figure out.

Cheapie's new player mod

The whitelist mod only players on the list can join.

My keyword interact mod

a fork of my keyword mod with slightly more advanced feature(not made by me)


That whitelist mod looks promising, but there's still the problem of new players contacting you to get their name added.

As for the others, I guess they'd work, but automated interact-granting mods don't prevent griefers with an intention to cause havoc (if the server doesn't have a protection mod).
Last edited by aix on Wed Aug 03, 2016 14:56, edited 1 time in total.
Lightning Server
Courage is not the absence of fear, but rather the judgment that something else is more important than one's fear.
-Ambrose Redmoon
 

CWz
Member
 
Posts: 196
Joined: Tue Dec 24, 2013 17:01
Location: Banana Land
GitHub: chaoswormz
IRC: CWz
In-game: CWz

Re: [Proposition] Trusting players

by CWz » Wed Aug 03, 2016 14:54

aix wrote:As for the others, I guess they'd work, but automated interact-granting mods don't prevent griefers with an intention to cause havoc (if the server doesn't have a protection mod).


I don't think there is such as thing as fool's proof way to prevent griefing. short of having congress passing down griefing is cybervandalism as a cybercrime.
 

aix
Member
 
Posts: 15
Joined: Fri Jul 22, 2016 18:52
IRC: aix
In-game: aix

Re: [Proposition] Trusting players

by aix » Wed Aug 03, 2016 14:57

CWz wrote:
aix wrote:As for the others, I guess they'd work, but automated interact-granting mods don't prevent griefers with an intention to cause havoc (if the server doesn't have a protection mod).


I don't think there is such as thing as fool's proof way to prevent griefing. short of having congress passing down griefing is cybervandalism as a cybercrime.


While that's true, people can judge other people a hell of a lot better than an automated system can, so I guess I'm looking for an automated system to facilitate the manual approval of users.
Lightning Server
Courage is not the absence of fear, but rather the judgment that something else is more important than one's fear.
-Ambrose Redmoon
 

User avatar
taikedz
Member
 
Posts: 676
Joined: Sun May 15, 2016 11:11
Location: Scotland, UK
GitHub: taikedz
IRC: DuCake
In-game: DuCake

Re: [Proposition] Trusting players

by taikedz » Wed Aug 03, 2016 15:57

I think what you're asking here is a way of picking out trustworthy players from swathes of untrustworthy ones.

Let's think about this in the wider "Internet" sense. What mechanisms are used there?

For example, what prevents me from logging on to an IRC channel as celeron55 when he's not there and causing confusion (the defter the communicator, the more mess can be made to be cleaned up....). Or sending out files in corporate mailing lists whilst spoofing the From header? Easy.

The "security" model of Minetest is equally lax (perhaps by design?)

So you can't trust names. What can you trust? You defer to an arbitrarily chosen authority.

Example: A game mod + server-side module that allows back-checking accounts.

The server retains a list of names, and hashes passwords with salts, provided by games servers that are configured with the trust mod. Simple lookup API allows mod to query the server.

The mod asks the user for their centralized password - if it matches against the server's entry, then the player is granted privs on that server.

Admins can add/remove trust/mistrust points to user names, other players can trust/mistrust players with a command. Mod can lookup periodically player trust status. If locally configured mod (game server level, as opposed to the centralized trust servuce) decides the that player has too much mistrust, their privs are revoked.

This implies that the person/team hosting the main trust server can be trusted to be fair and neutral. But anyone could run a server. For example, the MineTestForFun team could run one specifically for their group of servers.

If we could somehow decentralize the trust network such that admins can share with eachother that would be better.


But the short of it is, the login system of Minetest is open to the Net. It's not that we as admins don't trust our players - our players themselves can't trust or be given guarantees that other players are present in the same spirit. Protection mods are probably the sanest ways of allowing fair players to be minmally at risk from those who just want to watch our worlds burn.
 

aix
Member
 
Posts: 15
Joined: Fri Jul 22, 2016 18:52
IRC: aix
In-game: aix

Re: [Proposition] Trusting players

by aix » Wed Aug 03, 2016 17:10

I guess using a third party to authenticate players would be a good idea, but then you'd have to register on it before being allowed into the server, which could potentially deter actual users.

Here's a thought: let players into the server, and make them follow a maze of sorts, with the promise of getting the keyword to get interact, but at the end of the maze, instead of a keyword, put an alternative address and port, and a password, so players that have passed the maze can play on the actual server.
Lightning Server
Courage is not the absence of fear, but rather the judgment that something else is more important than one's fear.
-Ambrose Redmoon
 

User avatar
taikedz
Member
 
Posts: 676
Joined: Sun May 15, 2016 11:11
Location: Scotland, UK
GitHub: taikedz
IRC: DuCake
In-game: DuCake

Re: [Proposition] Trusting players

by taikedz » Wed Aug 03, 2016 18:10

aix wrote:I guess using a third party to authenticate players would be a good idea, but then you'd have to register on it before being allowed into the server, which could potentially deter actual users.


Fair point and i had thought as much. A way to make it seamless from within game would be key to its viability.

aix wrote:Here's a thought: let players into the server, and make them follow a maze of sorts, with the promise of getting the keyword to get interact, but at the end of the maze, instead of a keyword, put an alternative address and port, and a password, so players that have passed the maze can play on the actual server.


You underestimate griefers and that just strikes as an open invitation to them to give you the finger by griefing even harder :p And would surely deter real players too even more than simply registering...
 

aix
Member
 
Posts: 15
Joined: Fri Jul 22, 2016 18:52
IRC: aix
In-game: aix

Re: [Proposition] Trusting players

by aix » Wed Aug 03, 2016 19:33

A great way to catch out a griefer would be to grant them interact, and teleport them to a simulated village, with real-looking buildings, owned chests, some expensive materials, and no protection. If they dig a set number of blocks, their interact can be revoked, however good players will pass through the village, or perhaps be teleported out after a set amount of time.
Lightning Server
Courage is not the absence of fear, but rather the judgment that something else is more important than one's fear.
-Ambrose Redmoon
 

User avatar
taikedz
Member
 
Posts: 676
Joined: Sun May 15, 2016 11:11
Location: Scotland, UK
GitHub: taikedz
IRC: DuCake
In-game: DuCake

Re: [Proposition] Trusting players

by taikedz » Wed Aug 03, 2016 22:05

aix wrote:A great way to catch out a griefer would be to grant them interact, and teleport them to a simulated village, with real-looking buildings, owned chests, some expensive materials, and no protection. If they dig a set number of blocks, their interact can be revoked, however good players will pass through the village, or perhaps be teleported out after a set amount of time.


...... interesting....... i like this idea........

That being said it it will be quickly cottoned on to..... and griefers can circumvent by just leaving that area alone once they figure it out and use a new acct.....

Maybe actively identifying people truly building, and removing people who dig those areas too much....?

I think at this level it just makes more sense to make people protect their areas.

The TPS Xtreme server uses a maximum-two-areas-per-user system which i think is a good balance between protect freely from the start and preventing protection spamming.....
 

CuriousNoob
Member
 
Posts: 73
Joined: Wed Jan 27, 2016 09:43

Re: [Proposition] Trusting players

by CuriousNoob » Thu Aug 04, 2016 12:53

aix wrote:
CWz wrote:
aix wrote:As for the others, I guess they'd work, but automated interact-granting mods don't prevent griefers with an intention to cause havoc (if the server doesn't have a protection mod).


I don't think there is such as thing as fool's proof way to prevent griefing. short of having congress passing down griefing is cybervandalism as a cybercrime.


While that's true, people can judge other people a hell of a lot better than an automated system can, so I guess I'm looking for an automated system to facilitate the manual approval of users.



Hi again aix, :) ... this time in-forum rather than in-game ( jj_ )

Good thread. Tricky problem.

Cracking it could be fun --- it's actually a form of Games-Design in itself.


( TL;DR summary to follow :D )

Long version, thought-trains, comments :

First comes ideas, design, the concepts and strategy. Second is the material nuts and bolts of implementation. Often the latter shapes the former, sometimes limiting, sometimes revelatory and inspiring new directions ... like the relationship between Science and Engineering.

''Griefers'' is jargon I hadn't really encountered before Minetest. But it certainly fits.

There's a closely-related old expression ''fools and rogues.''

We want to sift them out, early, silently, automatically, along with whining high-maintenance children...

The reason I joined your new server was the ''UNIX Geeks'' tagline --- ongoing curiosity and identity-stuff.

The *nix crowd know the power of small single-purpose tools combined in pipelines. And of groups and permissions.

My immediate thoughts on your question are really extending and amplifying your Lightning server's spawn tower-trial philosophy. I'll flesh-out the musings with details and suggestions.

Whilst your headline issue is about trustworthiness, a huge part of the grief and annoyance seems to come just from the sheer immaturity of players --- an opinion distilled from seeing myriad forum-posts and in-game chatter.

And there's clearly a growing supply. I've noticed literally dozens of ''Sadie[[:digit:]]{4,}'' players popping up in worlds I've been visiting. Chatting/behaving in ways that speak volumes. I don't know the source, but my guess is it's some naff phone/pad app-store thing driving these casual fleeting visits.

Mods blacklisting certain types of obviously-algorithmic player-name would help, and I believe such mods are already out there on some servers. Their terse sign-up rejection-messages could be much better utilised however, providing an olive-branch of guidance for the genuine beginners.

Sadly, name-blocker mods offer only a tissue defence against the immature and untrustworthy.

Thinking about typical developmental patterns, the main mileage has to be in leveraging correlates of maturation : testing (directly or indirectly) for thresholds of Knowledge and Ability --- plus indicators of Temperament.

Most of the typical kiddlies we need to weed-out have not reached certain personal psycho-social levels (and indeed, looking at wider society, some never do).

This intellect- and age-related developmental-deficiency offers the most scope for testing, screening.

Your deliciously simple castle-tower test clearly requires threshold levels of 1) Reading and 2) I.Q. --- plus resourcefulness and motivation and enough in-game control skill.

There's a phrase I like which represents a fundamental of rational-thought and problem-solving. That phrase is ''Necessary and Sufficient.'' Most people haven't even heard of it, let alone properly comprehend it or its significance.

Your spawn-task principle is necessary, just not sufficient. It needs extending, compounding, in a series of sandboxed tests and tasks, demonstrating the required thresholds.

Priveleges definitely need to be granted in a staged way --- much like with children in real life.

Your non-Interact initial default is clearly the best defence.

But in practice of course, just being in the server-world with freedom to roam and freedom to chat allows the immature and irresponsible to be annoying in various ways which drain the time and emotional energy of admins and players.

Given how (mutually-)annoying it is for kiddlies to be hassling on the chat-stream but being ignored or scolded by existing players, I heartily agree with an earlier suggestion, I think there's a strong case for not even Shout priveleges initially.

Beyond mute on sign-up, an initial-location ''open-jail'' could limit any free-to-roam annoyances, preventing for example these kiddlies from tailing players and getting in the way psychologically. On E.S.M.C. they start players in a fully-protected space-platform, and require them to find and operate a travelnet down to spawn.

This space-platform idea can be extended, and has much to recommend it.

Starting in a jail-area location and unable to Shout or Interact would prevent a massive amount of trouble. While they're still geometrically-limited, impotent and mute, out-of-sight and out-of-mind, they become a zero-maintenance zero-annoyance irrelevance for existing players and admins.


''If all else fails, read the instructions...''

It was a few years ago I first stumbled across the use of '' tl;dr '' --- it stuck, not least because too often (as this post proves) I know ''mea culpa.''

These problem-player types typically don't --- and won't --- properly Read and Understand rules and instructions. In general, their Ability and Temperament are below the desired threshold.

Deliberately imposing a tl;dr element would heavily screen such players.

The trustworthy, more mature, responsible, conscientious and capable players won't mind this ''need-to-read'' during the screening-process if the reasons are clear and up-front as part of the induction/testing.

In the initial Captive Zone, we'd have not just one, but a series of in-game challenges like your castle-tower and much more.

In-game signs could require not only reading in-game directions, but also consulting out-of-game website-link information before being able to progress through the Captive-Zone challenges. To minimise solution-sharing hacks, the information might be dynamically- or even individually-generated, maybe from a pool of several static variants, whatever, anything from simple Captcha-esque passcodes through to Detailed Directions on completing a particular sub-task.

Make it all demanding enough and it will seriously test every prospective-player's mettle : Ability, Knowledge, and Temperament.

The casual pad-clicking kiddlies would simply fail and go (elsewhere), unseen, unheard, at no time troubling the wider server and its established players.

Staged granting of privileges as reward for solved challenges would make it obvious there's a progression into unrestricted membership of the server's world. I believe many Gamers are driven by ''cracking the next level.''

Present it like a ''12 Labours Of Hercules'' or some better-known and more contemporary redemption or rite-of-passage trope and it could actually make it quite attractive. And an enjoyable, intellectually-challenging game-design project for the server's regulars to design and build.

Key to implementing a lot of this ''Captive-Zone plus Challenges'' idea might be borrowing from the well-established strategy used by the whole *nix family : access-control, Users, Groups and Permissions. I believe Shared-Chests already allow multiple (named) players to use an otherwise-protected block, and we already have ''admin'' and ''moderator'' status in Minetest, so surely it shouldn't be much further to have proper *nix-like Groups and Permissions within the worlds.

The very first Captive-Zone challenge should obviously be just reading and moving intelligently to locate a deviously dangerously obfuscated exit-point, much like your castle-tower lava-hole, but moreso. With lots of signs to grind-down their tl;dr tolerance. ;)

The exit for each challenge would be the entry-point for the next, in a one-way system.

Sub-tasks could also double as a basic tutorial covering a number of the in-game techniques. I believe there was something like that on a SkyBlock world I visited. And if I'm honest, despite months since I discovered Minetest, it was only recently I myself finally learned the [Shift]-key Sneak trick for climbing. :D

Subsequent challenges would test to ever greater degrees the candidate's powers of reasoning and capacity to follow directions.

I have no direct personal experience placing them, nor which mods contain them, but I do know from the Craft-Guides and Spawns in other worlds that ''proximity-detector-blocks'' are available, as are ''activation buttons'' and dynamically-modifiable ''display-text'' ( like you put on my furnaces :D ).

Maybe also some derivative of the Mobs or AI-NPC type mods could allow dynamic puzzles and real-time modification of a candidate's task-environment. AI-Mob-blocks could move away from a player or toward, blocking or enabling access --- thus a Sokoban-style task becomes feasible despite not having Interact priveleges. I could even envisage too, in-game challenges like the game Picmi ( https : / / games.kde.org/game.php?game=picmi ), a distant cousin of the Sudoku family.

Automatic and dynamic blocks surely permit all sorts of devious and challenging tasks to be devised, including yet more tl;dr stress-testing. Separate and sandboxed from the main world, a mod could perform an automated reset of the task-zone map-blocks after the player exits.

Foreseeably there might need to be several identical task-zone map-blocks accessed from a spawn foyer waiting-room with multiple travelnet cubicles which are modded to include a ''Vacant'' or ''Occupied'' charateristic, thus coping with multiple simultaneous candidates.

Simply adding such an enforced delay, whether waiting to start the challenges or at any point thereafter, would see off those of the temperament who'd fail the ''Stanford Marshmallow'' and related tests.

Successfully completing all the Captive-Zone challenges could drop the player into an Audition-Zone, a much larger map sector but still effectively a jail-area, maybe with bedrock-esque mods, separate and well away from the main world and established players' builds.

In this Audition-Zone the new players would have Shout and Interact, and have a set time in which to demonstrate reasonable behaviour, maybe a month, knowing full well they're on probation, auditioning, needing to impress with mature chat and competent responsible builds. The Audition-Zone would be permeable for established players and admins to inspect progress and results.

Maybe some sort of ''Karma-Points'' or ''Kudos'' or ''Creds'' mod could display to these untrusted players an indication of approval-rating by established members --- a bit like Health-Point Hearts or Hunger-Point Loaves in the HUD.

The HUD could (mischievously) be made to look like Gold-Stars or De-Merits from school-days. :D

A GUI form-spec accessible only to established trusted players could allow voting-up or voting-down for instances of impressive or unacceptable behaviour or chat.

At the end of the audition/probation month, any score over a certain threshold could automatically confer normal priveleges and escape from the Audition-Zone to the wider world map, and slightly below that score granting them another month probation perhaps.

I'd better shut up for now. It's been more about design concept ideas than implementation, but hopefully some useful input somewhere within it.

I'll try to do a tl;dr summary ... although, to be honest, I'm not sure whether the ''tl;dr-crowd'' can actually be Trusted :D LOL

jj_

.
 

User avatar
taikedz
Member
 
Posts: 676
Joined: Sun May 15, 2016 11:11
Location: Scotland, UK
GitHub: taikedz
IRC: DuCake
In-game: DuCake

Re: [Proposition] Trusting players

by taikedz » Fri Aug 05, 2016 10:27

Hi jj_

Nice insightful post there :-)

However there are a couple of problems

1/ The whole trials-first approach is based on the idea that any legitimate player willo obviously be a highly literate and patient engineer/puzzle solving types and anyone who isn't is a potential griefer -- and even then, some highly literate engineers just come to places to play.

2/ The idea that griefers cannot be capable and deft puzzle solvers is incorrect. Sure you will catch petty vandals out, but you probably know from experience already that posing a challenge just invites more deft challengers - fair players and griefers alike. You might find things broken in /very/ interesting or even simply thorough ways once they get in.

3/ Since you refer to the UNIX philosophy of a task-focused tools in pipelines, I would like to point out that protection is that tool for the purpose of countering griefers, and that spawn puzzle mazes to "legitimize" players is a highly-complex, over-engineered way of going about this.

Of course, if you prefer that technique, if you own the server, there is no reason why you shouldn't implement it in that way. But you'll turn away legitimate players; and whilst you'll certainly attract the deft kindred players, you will get the level of griefers on equal par, and will still have to resort to protection mechanisms already extant.

4/ The only thing we have ended up discussing is pre-emptive measures. Protection mods act when griefers start griefing, not before.

aix's desire to unlist the server, my suggestion of a trust network, your engineered tower maze solution, and audition-build solution (a lot of moderation and evaluation overhead for the admin by the way!) are all pre-emptive measure to catch griefers before they enter the world.

This is very similar to, as I mentioned before, the wider Internet problem of sussing out the spammers, crackers and vandals on any forum, social network, reviews list etc.

Honestly, I'd rather see a vetting system that would allow players themselves to accumulate points on other players, and let an automated system take care of temporarily silencing shouty players, or banning altogether. Less overhead for the admin (who cannot approve/vet things whilst he/she is offline), and more distributed trust amongst the players.

Like a "/reprimand PLAYER" and "/praise PLAYER" adding points, and the automated system granting and revoking "shout" and "interact" dynamically for configurable periods of time.

In absence of these things -- protection and a set of moderators are the basic tools we have.
 

User avatar
taikedz
Member
 
Posts: 676
Joined: Sun May 15, 2016 11:11
Location: Scotland, UK
GitHub: taikedz
IRC: DuCake
In-game: DuCake

Re: [Proposition] Trusting players

by taikedz » Fri Aug 05, 2016 10:29

Also I hate games that "force" you into a tutorial mode.

I also hate applications that forcibily display tutorial popups in your face on fist run.

It's a sure way to turn away users, but maybe such games and applications don't want the likes of me :-)
 

aix
Member
 
Posts: 15
Joined: Fri Jul 22, 2016 18:52
IRC: aix
In-game: aix

Re: [Proposition] Trusting players

by aix » Fri Aug 05, 2016 12:14

CuriousNoob wrote:Hi again aix, :) ... this time in-forum rather than in-game ( jj_ )
[...]
jj_


Hi!
Just before you made this post (I think), I implemented various safeguards into the server I run to further the protection against griefers.
You correctly pointed out the ways players without interact can still hassle established users; for instance in chat or just by getting in the way. To combat these, I have moved the "trial" section about 100 blocks under spawn, and made it into a maze, which users have to complete with no privileges (except nointeract, which is necessary for the keyword mod). I tried to implement simultaneous tasks which require more than memory. Instead of making users read the rules then answer questions about them, I made the users guess the rules. Only players with the right moral code can get in, there's no easy way to get through -- for instance from memory, but this won't stop the good guys, for whom I hope the tests will be quite easy.
Essentially, this new test means that instead of the players following the rules given, the rules follow the players: the players make their own rules. Well, to a point, if they take a wrong turn in the maze, then their morals do not fit into this server's rules, and they'll be stuck forever (It's a one way system).
As I mentioned before, I did implement a memory test -- players had to remember a number given to them at the start of the maze. Remembering this number after the distraction of the maze means only users with a certain working memory will be allowed in.
I have thought about - but elected not to incorporate - a jumping puzzle into the maze. Not only would this have completely screwed over mobile users, but it would have demanded a certain level of expertise from the players.
Another thought I had was putting a maze (a proper maze) into the test, to evaluate the users' skills in navigation and orienteering.

I don't think that vetting players on the server for a month would be the way to go, the maze I implemented has, so far, blocked 100% of troublemakers that have joined. Vetting users for a month would not only deter the skilled players, some of which just want to build. However, I do think you're on the right track there. By letting the players elect moderators and administrators, we can help to establish not only a self-governing system, but one build on trust and respect, much like the democratic system we have today (well, perhaps with a bit less politics and well.. bullshit).

Regards,
IBM AIX
Lightning Server
Courage is not the absence of fear, but rather the judgment that something else is more important than one's fear.
-Ambrose Redmoon
 

CuriousNoob
Member
 
Posts: 73
Joined: Wed Jan 27, 2016 09:43

Re: [Proposition] Trusting players

by CuriousNoob » Sat Aug 06, 2016 17:23

talkedz...

aix...



Whoops! I guess I misunderstood how (not-so-) high we wanted the bar to be set. LOL :D

But a couple of clarifications are worth adding.

Truth be known, I was only just getting started on ideas and the rationale behind them. But I'm very slow at writing and deliberately called time on that first post. Eeek! LOL :D


Non-negotiable :

''Trust'' is a multi-facetted thing which can only emerge after sufficient time and contexts have played out.


@talkedz

Absolutely, I totally agree, the ''bad-guys'' can be extremely clever too. Bright but anti-social individuals ''gaming the system'' for reasons of gain and ego is exactly why so much of real-life contemporary economics and society looks the way it does, sadly.

Hence my suggesting an overkill ''belt-and-braces'' approach, stress-testing crucially for Temperament and Maturity, in that deliberately protracted, probationary, staged way.

I suppose I was responding primarily with the stated-aims posted by aix in mind --- those when he originally announced Lightning, at least in the first and second incarnations (the current shortened version no longer elaborates on server philosophy and goals).

Those initial aims, iirc, did seem to be : adult bias, detail-focused, collaborative, unixy, geeky.

I obviously seriously underestimated the desire to remain widely accessible and inclusive.

Ironically, I actually thought I'd been including concessions in that direction by suggesting the tutorial element in the tasks for ''genuine beginners'' of the right type who might need it.

Ignorance is the default initial state for us all. It wasn't that many months ago I was in that ''beginner and also genuine'' category myself and welcomed every bit of helpful guidance, which included in-game tutorial elements here and there. I'm long the type that silently thrashes StartPage for wiki and forum answers rather than trouble anyone else with my newbie questions, but as with a lot in the FOSS ecosystem, info can be hair-pullingly patchy and/or out-of-date.

When you ''don't know what you don't know'' and when systems rest on arbitrary protocols and procedures --- which can't all be deduced or second-guessed or found by experimentation among the myriad combinations and permutations of contexts and keystrokes and clicks --- it's a Catch-22 condemnation to eternal noobness unless you get to be apprentice with proficient practitioners or have proper documentation available.

Probably I should have been clearer that in suggesting educative tasks, challenges with a tutorial element, it did not imply ''in-your-face'' enforced-classroom modes nor pop-up ''spoon-feeding and hand-holding'' --- an approach which would undermine (no pun intended :D ) the broader IQ-screening for entry.

Genuine question : how much of a target for alpha-geek griefers are these voxel-worlds..?

Enough to endure such time-consuming probation tasks and delays with reduced privileges..?

Are there any reliable stats, even any anecdotes?

Would not ''hacked clients'' be more their weapon of choice anyway? And surely hacked-clients is a very different topic than Trustworthiness of players as per O.P. above?

The erstwhile paragon Xanadu had no real entry-screening it seems, and apparently still hasn't recovered since a script-kiddy wreaked havoc. Old players and potential new players are still suffering the fallout!

Venn-diagrams ...

the sufficiently IQ-puzzle-capable

AND sufficiently able to hoodwink established players with a non-slipping facade of emotionally-mature reasonable temperament throughout a protracted probation

AND sufficiently motivated to endure all that just to cause whatever extent of griefing

... surely tiny, vanishingly-small, or am I being naive about griefer mentality and ubiquity?

Looking for analogues in the real world, how do prestigious clubs and societies approach the new-members issue?

All-too-aware of Gaussian distributions, sure, I freely admit very few would pass my suggested arrangement; they'd be mainly patient thoughtful mature creative and tecchie-types. People like us, people like me. No big surprise, I've suffered and lost too much in real life at the hands of the other types. I'm not overly fond of a lot of them. What's more, there's already plenty of existing opportunities for the neuro-typicals, the mainstreamers, the ordinaries, to get their quick fixes of dumbed-down instant-gratification play elsewhere, surely..?

I'll say again. IIRC : adult bias, detail-focused, collaborative, unixy, geeky.

But yes, of course, most of those who'd fail tough tests aren't (intentional) griefers or untrustworthy.

But being able to ''trust'' that an individual's intentions and behaviour will be considerate, ethical, and constructive and not be selfish, malicious and destructive, is not the same thing as being able to ''rely on'' their results always being benign.

Not everyone is sufficiently sensitive to the ripples and wake and footprint they leave.

There's a reason the old expression includes ''fools'' alongside ''rogues.''

And a reason that sanity usually legislates children away from vehicles and weapons.

Immature or lower-ability does not mean a Bad Person! It does however skew certain probabilities and require more from the environment --- better pre-emptive design, more robustness, more monitoring, more chemspill, more repair, whatever. Have you ever taught or cared for Children or Special-Needs? Probabilities vary across the spectrum.

Anyway, returning to the core theme of Trust and Griefing...

Yes, definitely Owner/Area Protections-systems do need to be retained ... but I was hoping to explain how they might be augmented and enhanced.

I already mentioned multi-user Shared-Chests; I do still think some improved form of in-game Groups and Group-Permissions would be the right direction to pursue, enhancing both gameplay and anti-griefing.

Relevant reference : from my very first post on these forums I learned (surprised and miffed to find) that not all items/blocks could be crafted by all players. Specifically I was trying to craft a teleporter, but was told that only admins could do that in that world.

Based on that discovery, I've wondered whether, for example, all block-types might carry extra characteristics --- comparable to the ''diggable-by-minimum-pickaxe-type'' --- characteristics which limit the block's diggability, placeability, craftability and usability to certain threshold group membership, in concert with a much greater number of ranks than just player/moderator/admin.

Imagine, say, 1024 binary entries in a privileges-matrix, thought of as a 16 x 64 array, 16 levels of increasing rank able (or not) to do certain things with the 64 groups of blocks and items --- like dig, pick-up, own, craft, use, place-temporary, place-permanent, place-more-than-N, can-jump-Z, can-swim-W etc etc etc whatever whatever for any of the timber, stone, ore-group-A, ore-group-B, tool-group-P, tool-group-Q, bucket, fire, lava, teleporter, machine, etc etc etc.

The 16 groups might be strictly cumulative or else a ''pick'n'mix'' way of adding diversity in abilities and limitations.

Diversity itself would tend to enrich and favour collaborative gameplay, teams of complementary individuals being needed to complete a project. Like chefs need farmers and plumbers and electricians. Or dwarfs need hobbits. Or warriors need medics.

Those 16 groups would be a shorthand way of assigning multiple privileges. And it needn't be just 16 or 64.

Not all need be used in every world, but each server-owner could name and delegate them locally in a way which reflects the nature of the world they're trying to create. A novice, a student, an apprentice, a woodworker, a craftsman, a master-craftsman. A hobbit, a dwarf, an elf, a wizard. Private, corporal, sergeant, captain. Labourer, miner, toolmaker, engineer, terra-former ... deity! :D

You get the idea I'm sure.

And instead of just picking textures/skins a new player, or second-promotion novice, might actually choose a particular character-trajectory in the world with a different mix of permissions-abilities than the next player. Engineer rather than warrior. The experience would differ.

I believe games-designers refer to ''depth'' --- and from what I've experienced personally in these few months of MT, there are times and worlds where you'd definitely want more depth. I think that's one contributory factor why hostile mobs (AI mobiles) seem popular, it introduces unpredictable complexity and an element of mild risk. PvP on the other hand... meh, sometimes just another form of griefing, tedious and wearisome, rewarding whoever happened to have signed up first and got to the best weapons and armour first (like in real life). A number of worlds seem to have ''no PvP at spawn'' rules for obvious reasons.

Sorry. Buzz! ''Deviation!'' Gotten a bit carried away off-topic. :D

Back to *nix-like Permissions and Trust :

Granularity of Control / Staged Loosing of Restrictions

The core idea of expanded enhanced group-permissions privileges could permit finer-grained admin-control over what the new and untrusted can actually get up to, universally, not dependent on the presence of explicitly Protected Areas.

Example :

If...

a maze-solving untrusted newcomer, assigned to say ''novice-carpenter'' group, has only group-permissions of :

Code: Select all
shout, dig, pick-up, own, craft, temporary-place (with some sort of block-characteristic:expiry_date:YYMMDDHHMMSS), dig-no-more-than-N, place-no-more-than-P, whatever whatever, etc etc


applicable to only blocks/items of group-types :

Code: Select all
wood/timber, foodstuffs, dirt, sand, stone, coal, steel, chest, furnace, whatever whatever etc etc


and stays in that ''novice'' group until they've earned enough ''Creds'' Approvals-Points from established players and admins to be promoted,

then...

that newcomer intrinsically cannot grief as much as quickly as they could under the present all-or-nothing system of instant Interact.

I suspect it's essentially the current /privs model but massively widened, .

Your ''/reprimand PLAYER'' and ''/praise PLAYER'' is exactly the sort of thing I meant (just CLI).

These approval-rating ''Creds'' above a certain level after a certain time-period would automatically confer a group-promotion (or choice of promotion), or else penalties for transgressing.

That's broadly what I had in mind when suggesting player-votable ''Karma-Points.''

I do believe actual humans need to do it though, not algorithms.

It's only over sufficient active Time with sufficient meaningful observation and interaction that Temperament and ultimately Trust can even hope to be assessed.

Within that period sensible restrictions/permissions need to be in place throughout.

Maybe I'm being naive again. Maybe the above permissions-model is way beyond the reach of mod-coders and firmly a minetest-devs core programme-design thing.

But I still do think it's a goal worth pursuing.

Sheesh, another tl;dr essay. :D

I'll stop for now.

---

@aix

Rather intrigued by your description of the new beefed-up spawn-trial. Might have to go have a sneaky peek. :D

Also, started looking at that ''MinePass'' thing. No idea what's under the bonnet, behind the scenes, nor how it's supposed to be used. But I've signed up anyway. Ever curious. :D

I suppose Bitcoin-esque ''blockchain-technology'' strategies might be able to make Trust gained in one world portable elsewhere ... although of course, there's a big difference between :

a) - certifying that a particular transaction has happened and
b) - certifying that an individual is worthy of Trust.

I believe this latter is still a fundamental problem in encryption-certificate circles and the notion of a Web of Trust, when at it's root, the only person you know you can trust is, well ... You!

---

Hope something helps somebody...

jj_

.
 

aix
Member
 
Posts: 15
Joined: Fri Jul 22, 2016 18:52
IRC: aix
In-game: aix

Re: [Proposition] Trusting players

by aix » Sat Aug 06, 2016 22:01

If humans run the validation system, then isn't what you're proposing a simple server without interact by default where moderators grant interact?
Such a system would be highly effective, but it'd require moderators, trusted players, and let's not even get into how we're going to elect those :P
Another idea is to gain money with in-game time, and actually BUY the further privileges. I think there's a limit to the amount of time a griefer is willing to spend, anyhow.
Lightning Server
Courage is not the absence of fear, but rather the judgment that something else is more important than one's fear.
-Ambrose Redmoon
 

CuriousNoob
Member
 
Posts: 73
Joined: Wed Jan 27, 2016 09:43

Re: [Proposition] Trusting players

by CuriousNoob » Sun Aug 07, 2016 10:48

aix wrote:. . . I think there's a limit to the amount of time a griefer is willing to spend, anyhow.



Yep. Total agreement; it's the ''temperament and motivation'' thing.

Still, ''a thousand experiments can't prove a theory right, but just one can prove it wrong.''

So I'm interested to see if talkedz or others do offer any stats or anecdotes.
.
 

CuriousNoob
Member
 
Posts: 73
Joined: Wed Jan 27, 2016 09:43

Re: [Proposition] Trusting players

by CuriousNoob » Sun Aug 07, 2016 11:04

aix wrote:If humans run the validation system, then isn't what you're proposing a simple server without interact by default where moderators grant interact?
Such a system would be highly effective, but it'd require moderators, trusted players, and let's not even get into how we're going to elect those :P



Only partially, and only initially.

But it desperately needs those two things I suggested :

1) - a better, finer-grained interact-permissions progression, a matrix, not the present all-or-nothing /interact model, and

2) - more ranks than merely the current player-moderator-admin trio.

My assumption being that you and other server-owners do actually inhabit the worlds you create ... at least for a time after public-release.

A period enough to personally gauge --- because you're a human and smarter than any automated algorithms, well, at least until The Singularity renders us all irrelevant slaves and fugitives, :D LOL --- personally gauge the trustworthiness of the first permissions-limited individuals who first join your new world, enough to confer a form of ''Voting-Privilege'' which both talkedz and I were suggesting. From there, the whole in-game voting approval-ratings karma-points thing, the 'community-democracy' you mentioned earlier, would roll it onward.

It would make collaborative cooperative mature sustained player-to-player communication and interaction a necessity --- which matches how I interpreted your originally-stated philosophy and goals for Lightning.

Rousseau's Social Contract discussed governance, and there's much to recommend ''benevolent dictatorship'' --- which is, after all, what we're used to in *nix-land.

The all-powerful ''root'' is benign, and you as server-owner/admin are root.

Your MinePass suggestion, or something similar, could indeed help reduce timescales through reputation-portability, but it does of course add extra layers, and vulnerably external layers at that. A number of times at attempted sign-up MinePass was giving server/network errors. Also, the moment such a system starts any sort of compulsory metadata overreach, it will be actively shunned by those at the privacy/freedom-activist end of the FOSS-spectrum, potentially losing the constructive-feedback contributions of some very bright individuals.

If my initial assumption is incorrect and, instead, you or other server-owners are wanting to crank-out worlds in quantity but not actually inhabit them through their set-up period, nor actively monitor and moderate, then yes, the Trust issue does become much trickier.

Lord Of The Flies..?

If you're looking to delegate and semi-automate, at some point in the process, you as owner/root must trust someone/something somewhere, in-game or externally. Which surely requires you to have observed for sufficient time in sufficient contexts the speech and behaviours you consider indicative of trustworthiness.

Personally I don't think that role can ever be fully algorithmic.

As I mentioned, this apparently shares much with the encryption-community's issues. Webs-of-Trust and Key-Signing and Certifying-Authorities ultimately comes back one thing : whom do * I * trust, and for what..?

.
 

CuriousNoob
Member
 
Posts: 73
Joined: Wed Jan 27, 2016 09:43

Re: [Proposition] Trusting players

by CuriousNoob » Sun Aug 07, 2016 13:16

aix wrote:Another idea is to gain money with in-game time, and actually BUY the further privileges . . .


The idea of points or money is interesting ... although I can foresee how it could be abused, like in real life.

Would payments merely reflect time logged into a world?

Or would it have an interrogative element with responses proving genuine at-keyboard eyes/minds/fingers?

Or would payments only be for actually doing something, tasks completed, locations reached...?

Combination/permutation?

.
 

Byakuren
Member
 
Posts: 816
Joined: Tue Apr 14, 2015 01:59
GitHub: raymoo
IRC: Hijiri
In-game: Raymoo + Clownpiece

Re: [Proposition] Trusting players

by Byakuren » Mon Aug 08, 2016 00:14

clearly privileges should be a microtransaction, like on some of those minecraft "donation" servers
Every time a mod API is left undocumented, a koala dies.
 

twoelk
Member
 
Posts: 1280
Joined: Fri Apr 19, 2013 16:19
Location: northern Germany
GitHub: twoelk
IRC: twoelk
In-game: twoelk

Re: [Proposition] Trusting players

by twoelk » Mon Aug 08, 2016 03:24

interesting thread with some fascinating statements and ideas -

LazyJ of the Wazuland server once wrote an interesting characterization about players on public servers. I guess not much has changed except that we now also have the kids on mobile devices that have difficulties controlling the avatar and start griefing by accident and stick to doing that as everything else is more difficult.

Indeed the problem is not new. Some may remember the old GameBoom server that seems to have been griefed to death. I can't remember wether the counter measures proposed later ever worked out but some of them do sound familiar to things mentioned in thís thread.

I have played on many servers and have opened some of my maps for public access now and then and have learned that the only system that works really well is an alert team of human moderators that use a common code of conduct and apply a control system several steps deep. Often a cool down time is more effective than a quick total ban as many hot blooded players or others that like dangerous experiments can over time become valuable for the server or even minetest later on.

For example Dan Duncomb loved to play with TNT or do other stuff that most might consider as griefing, yet a lot of the mods he made or added to are still used. When I played on VanessaE's creative server a lot a young boy joined that only spoke german. He soon settled near my area and started exploring the possabilites of minetest. He was on the brink of being banned several times for experiments with lava getting out of hand and not understanding what the admins where asking him. A lot of my building in our shared area was my response to cope with some of his accidents. Yet he built interesting stuff and got better all the time and less destructive as he started to understand minetest better.

I can't remember who said it first but "Never attribute to malice that which is adequately explained by stupidity" (or ignorance or maybe bad GUI design for android devices in our case) is something we should never leave out of account. And people do learn and some do change to the better.

Several servers have used several interesting ways to cope with annoying people. Mauvebic for example used to teleport people far away to a special area. The game.d-l-k.net server had a newbie area and if players proved worthy builders would allow them to other mostly themed areas, besides having a little maze at spawn that included some jumping skills. Often the spawn would be crowded with dozens of players though that blocked each others views and made the jumping difficult for even experienced players on pc.

lag01 wrote the Kick voting mod for his "just test" server that did not really have the desired results just as the self governing didn't really work when the "good players" where not online.

Monitoring a server well can be really time consuming and cause a lot of headaches not related to technical issues. I guess this is the part that most people forget when they create a server.
 

aix
Member
 
Posts: 15
Joined: Fri Jul 22, 2016 18:52
IRC: aix
In-game: aix

Re: [Proposition] Trusting players

by aix » Mon Aug 08, 2016 11:26

I think that an "experience" system would benefit servers greatly. This system would upgrade the ranks of ordinary users so they are capable of more. Perhaps instead of the ability to place certain block, an upgrade could permit a user to build in a certain "elite" area. Think about it, everyone's rank dependent on their building skill and fairness. I know that machines cannot fairly decide if someone is trustworthy.
That's where voting comes in. There can be elections! In these elections, only people with a certain rank can attend. There can be elections at different ranks. The assumption here is that the players who spend more time on the server and as a result get to be of a higher rank are well acquainted with the candidates in the election and bear at least some traits of a conscious, decision making being. This would bring together automation and human intuition to create a true self-governing server. Even the founders of the server could have limited access, so it would be possible to literally "build your way to the top", rising through the ranks, election after election, and essentially take the seat of the founder of the server, if you prove to be better at their task than them.

Taking the less violent approach, administrators and moderators can simply accumulate instead of having to actively participate in the server to retain their rank, however I feel that this approach would be less beneficial to the server, since it would not actively encourage participation, once they reach a certain status, they can just slack off. However, granting privileges at threat of taking them away if the candidate becomes less well suited is an example of negative reinforcement. Positive reinforcement would be granting privileges at threat of not granting any more if the candidate slacks off. This would cause less havoc, and perhaps reduce the probability of rogue operators.

One problem with this system is that there would have to be a ladder with a rather large number of steps, or there would have to be a long wait in between upgrades in rank. Either way, I think that there should be an unlimited number of positions open at every rank, but that there should be restrictions on what even the administrators can and cannot do. Once a player reaches administrator, I'm guessing they'll have the server priv, which would mean they could shut down the server if they wanted. If they've reached the position, I'm guessing they're trustworthy and won't do this, unless they have a good reason, but instead of not granting the server privilege, I think you could delay granting it until the rank of super-administrator, which would essentially give the player shell access to the server's box. A huge remote hole, I'm sure, but with all the appropriate restrictions in place, this system should be fair and safe.
Lightning Server
Courage is not the absence of fear, but rather the judgment that something else is more important than one's fear.
-Ambrose Redmoon
 

Next

Return to Servers



Who is online

Users browsing this forum: No registered users and 6 guests